[Info-vax] Final Orace release on VMS.

[Arne Vajhøj]

On 11/16/2020 1:08 AM, Dave Froble wrote:
> On 11/15/2020 6:49 PM, Arne Vajhøj wrote:
>> On 11/15/2020 10:22 AM, Dave Froble wrote:
>>> On 11/15/2020 8:22 AM, Marc Van Dyck wrote:
>>>> If we need to dedicate significant amounts of
>>>> resources to develop something just to keep the OpenVMS platform in
>>>> business, the decision will certainly be to dedicate those resources
>>>> to eliminate that platform from our IT landscape once for all.
>>>
>>> What I mentioned is what I commonly refer to as a relay.  Basically a
>>> communications application.  I've implemented some to get between VMS
>>> applications and communications with services available on the
>>> internet, mainly due to the shortcomings of the TCP/IP and SSL/TLS
>>> available on VMS.
>>>
>>> Obviously the design of such an appliance will depend heavily upon the
>>> required communications.  But "significant amounts of resources" is
>>> not in that picture.  Something like that should take much less than a
>>> single man year, and you mentioned the alternative is perhaps 200 man
>>> years.
>>>
>>> As a possible incentive, I have no idea of the cost of Oracle client
>>> for VMS including support, but such a relay would not incur such
>>> costs, perhaps actually saving some costs, and that would be yearly
>>> savings vs a single development cost.
>>
>> It is not just the relay application that need to be developed.
>>
>> They would also need to change all their applications from using
>> Oracle client (whether embedded SQL or direct OCI calls) to interact
>> with the relay. And that would be rather intrusive.
>>
>> And depending on what features are needed then the relay
>> application may be more complicated. I would be a bit
>> concerned about the transactional integrity.
>>
>> It can be solved. Oracle solved it with the way they
>> support thin JDBC driver for Rdb. But most companies
>> want to focus their development effort on the business
>> problem and not on developing let us call it platform
>> components.
> 
> 200 man years can be a rather telling incentive ....
> 
> Besides, I like developing "platform components" and "system software".

Lots of developers prefer coding interesting technical stuff over
rather trivial business rules.

But CIO's tend to be skeptical.

Arne